全球知名图形处理单元（GPU）制造商英伟达近日宣布，正式成立了一个全新的研究部门——GEAR，全称为通用具身智能体研究。这一举措标志着英伟达在人工智能和虚拟现实领域的探索迈出了重要一步。GEAR部门将由AI领域的资深专家，高级研究科学家Jim Fan以及Yuke Zhu教授共同领导，他们的目标是构建能够在虚拟与现实世界中无缝切换的全能智能体。
据英伟达透露，GEAR的核心任务是研发能够适应多模态环境和各种场景的智能体基础模型。这意味着，该部门的工作将不仅限于传统的计算领域，而是要打破界限，实现智能体在虚拟世界和物理世界中的广泛运用。研究团队将主要聚焦四大关键研究领域：
1. **多模态基础模型**：旨在创建能够理解和处理包括视觉、听觉、触觉等多种信息输入的智能模型。
2. **通用型机器人研究**：探索智能体在现实世界中的应用，如开发能够执行各种任务的机器人。
3. **虚拟世界中的基础智能体**：在数字环境中创建具有自主学习和适应能力的智能实体。
4. **模拟与合成数据技术**：利用模拟和合成数据来训练智能体，提高其在复杂环境中的性能和适应性。
这一系列研究方向预示着英伟达正致力于构建一个更为真实、智能且相互连接的未来世界，其成果将可能深远影响人工智能、机器人技术以及虚拟现实的发展。通过GEAR部门，英伟达不仅展示了其在科技创新上的雄心，也进一步巩固了其在全球AI和图形处理领域的领先地位。
英语如下：
**News Title:** “NVIDIA Launches GEAR Research Division to Create Cross-Reality Intelligent Agents”
**Keywords:** NVIDIA, GEAR Division, Intelligent Agent Research
**News Content:**
Renowned graphics processing unit (GPU) manufacturer NVIDIA has recently announced the establishment of a new research division called GEAR, standing for General Embodied AI Research. This move signifies a significant step forward in NVIDIA’s exploration of artificial intelligence (AI) and virtual reality. The GEAR division will be co-led by seasoned AI expert and senior research scientist Jim Fan, along with Professor Yuke Zhu, a prominent figure in the field.
According to NVIDIA, GEAR’s core mission is to develop foundational models for intelligent agents capable of adapting to multimodal environments and diverse scenarios. This implies that the department’s work will transcend traditional computing boundaries, enabling widespread application of intelligent agents in both virtual and physical worlds. The research team will focus on four key areas:
1. **Multimodal Foundation Models:** Aimed at creating intelligent models that can understand and process inputs from visual, auditory, and tactile modalities.
2. **General-Purpose Robotics Research:** Exploring the application of intelligent agents in the real world, such as developing robots capable of performing various tasks.
3. **Foundational Intelligent Agents in Virtual Worlds:** Creating autonomous, learning, and adaptive entities within digital environments.
4. **Simulation and Synthesized Data Techniques:** Utilizing simulation and synthesized data to train intelligent agents, enhancing their performance and adaptability in complex settings.
These research directions indicate NVIDIA’s commitment to building a more realistic, intelligent, and interconnected future. The outcomes of these efforts could have profound implications for the development of AI, robotics, and virtual reality. Through the GEAR division, NVIDIA not only showcases its ambition in technological innovation but also solidifies its leading position in the global AI and graphics processing landscape.
【来源】https://mp.weixin.qq.com/s/-ep_nCt6moygQEm0MjOxqw
Views: 0